Fuzzy dynamic clustering algorithm

Abstract

A three-stage dynamic fuzzy clustering algorithm consisting of initial partitioning, a sequence of updating and merging by optimisation of a characterisation function based on measures of fuzziness in a set is described. Unlike the conventional detection of disjoint initial clusters, the algorithm can extract overlapping initial cluster boundaries when the feature space has ill-defined regions. The membership function in Rⁿ involves the density of patterns at a point in addition to its Euclidean distance. The merging criterion involves the number of samples and the amount of fuzziness in the intersection of two clusters, and the disparity in their size. The effectiveness of the algorithm is demonstrated on the speech recognition problem.